def test_digits_modular_sparse():
model = MaxCoverageSelection(100, optimizer='modular', random_state=0)
model.fit(X_digits_sparse)
assert_array_equal(model.ranking, digits_modular_ranking)
assert_array_almost_equal(model.gains, digits_modular_gains, 4)
assert_array_almost_equal(model.subset,
X_digits_sparse[model.ranking].toarray())
def test_digits_naive():
return
model = MaxCoverageSelection(100, optimizer='naive')
model.fit(X_digits_cupy)
assert_array_equal(model.ranking, digits_ranking)
assert_array_equal(model.ranking, digits_ranking)
assert_array_almost_equal(model.gains, digits_gains, 4)
def test_digits_two_stage_sparse():
model = MaxCoverageSelection(100, optimizer='two-stage')
model.fit(X_digits_sparse)
assert_array_equal(model.ranking[:4], digits_ranking[:4])
assert_array_almost_equal(model.gains[:4], digits_gains[:4], 4)
assert_array_almost_equal(model.subset,
X_digits_sparse[model.ranking].toarray())
def test_digits_approximate_sparse():
model = MaxCoverageSelection(100, optimizer='approximate-lazy')
model.fit(X_digits_sparse)
assert_array_equal(model.ranking, digits_approx_ranking)
assert_array_almost_equal(model.gains, digits_approx_gains, 4)
assert_array_almost_equal(model.subset,
X_digits_sparse[model.ranking].toarray())
def test_digits_stochastic_object():
model = MaxCoverageSelection(100,
optimizer=StochasticGreedy(random_state=0))
model.fit(X_digits)
assert_array_equal(model.ranking, digits_stochastic_ranking)
assert_array_almost_equal(model.gains, digits_stochastic_gains, 4)
assert_array_almost_equal(model.subset, X_digits[model.ranking])
def test_digits_two_stage_init():
model = MaxCoverageSelection(100,
optimizer='two-stage',
initial_subset=digits_ranking[:5])
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:10], digits_ranking[5:15])
assert_array_almost_equal(model.gains[:10], digits_gains[5:15], 4)
def test_digits_naive():
model = MaxCoverageSelection(100, optimizer='naive')
model.fit(X_digits)
assert_array_equal(model.ranking[:15], digits_ranking[:15])
assert_array_equal(model.ranking[:15], digits_ranking[:15])
assert_array_almost_equal(model.gains[:15], digits_gains[:15], 4)
assert_array_almost_equal(model.subset, X_digits[model.ranking])
def test_digits_lazy_init():
model = MaxCoverageSelection(100,
optimizer='lazy',
initial_subset=digits_ranking[:5])
model.fit(X_digits)
assert_array_equal(model.ranking[:5], digits_ranking[5:10])
assert_array_almost_equal(model.gains[:5], digits_gains[5:10], 4)
assert_array_almost_equal(model.subset, X_digits[model.ranking])
def test_digits_greedi_nl_object():
model = MaxCoverageSelection(100,
optimizer=GreeDi(optimizer1='naive',
optimizer2='lazy',
random_state=0))
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:2], digits_ranking[:2])
assert_array_almost_equal(model.gains[:2], digits_gains[:2], 4)
def test_digits_greedi_ll():
model = MaxCoverageSelection(100,
optimizer='greedi',
optimizer_kwds={
'optimizer1': 'lazy',
'optimizer2': 'lazy'
},
random_state=0)
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:30], digits_greedi_ranking[:30])
assert_array_almost_equal(model.gains[:30], digits_greedi_gains[:30], 4)
def test_digits_greedi_nl_sparse():
model = MaxCoverageSelection(100,
optimizer='greedi',
optimizer_kwds={
'optimizer1': 'naive',
'optimizer2': 'lazy'
},
random_state=0)
model.fit(X_digits_sparse)
assert_array_equal(model.ranking[:2], digits_ranking[:2])
assert_array_almost_equal(model.gains[:2], digits_gains[:2], 4)
def test_digits_sieve_minibatch_sparse():
model = MaxCoverageSelection(100, random_state=0)
model.partial_fit(X_digits_sparse[:50])
model.partial_fit(X_digits_sparse[50:150])
model.partial_fit(X_digits_sparse[150:])
assert_array_equal(model.ranking, digits_sieve_ranking)
assert_array_almost_equal(model.gains, digits_sieve_gains, 4)
assert_array_almost_equal(model.subset, X_digits[model.ranking])
def test_digits_sample():
model = MaxCoverageSelection(100, optimizer='sample', random_state=0)
model.fit(X_digits_sparse)
assert_array_equal(model.ranking, digits_sample_ranking)
assert_array_almost_equal(model.gains, digits_sample_gains, 4)
def test_digits_lazy_sparse():
model = MaxCoverageSelection(100, optimizer='lazy')
model.fit(X_digits_sparse)
assert_array_equal(model.ranking[:3], digits_ranking[:3])
assert_array_almost_equal(model.gains[:3], digits_gains[:3], 4)
def test_digits_modular_object():
model = MaxCoverageSelection(100, optimizer=ModularGreedy(random_state=0))
model.fit(X_digits_cupy)
assert_array_equal(model.ranking, digits_modular_ranking)
assert_array_almost_equal(model.gains, digits_modular_gains, 4)
def test_digits_approximate_object():
model = MaxCoverageSelection(100, optimizer=ApproximateLazyGreedy())
model.fit(X_digits_cupy)
assert_array_equal(model.ranking, digits_approx_ranking)
assert_array_almost_equal(model.gains, digits_approx_gains, 4)
def test_digits_two_stage_object():
model = MaxCoverageSelection(100, optimizer=TwoStageGreedy())
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:4], digits_ranking[:4])
assert_array_almost_equal(model.gains[:4], digits_gains[:4], 4)
def test_digits_lazy_object():
model = MaxCoverageSelection(100, optimizer=LazyGreedy())
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:3], digits_ranking[:3])
assert_array_almost_equal(model.gains[:3], digits_gains[:3], 4)
def test_digits_stochastic():
model = MaxCoverageSelection(100, optimizer='stochastic', random_state=0)
model.fit(X_digits_cupy)
assert_array_equal(model.ranking, digits_stochastic_ranking)
assert_array_almost_equal(model.gains, digits_stochastic_gains, 4)
def test_digits_approximate():
model = MaxCoverageSelection(100, optimizer='approximate-lazy')
model.fit(X_digits_cupy)
assert_array_equal(model.ranking, digits_approx_ranking)
assert_array_almost_equal(model.gains, digits_approx_gains, 4)
def test_digits_naive_object():
model = MaxCoverageSelection(100, optimizer=NaiveGreedy())
model.fit(X_digits)
assert_array_equal(model.ranking[:4], digits_ranking[:4])
assert_array_almost_equal(model.gains[:4], digits_gains[:4], 4)
assert_array_almost_equal(model.subset, X_digits[model.ranking])
def test_digits_two_stage():
model = MaxCoverageSelection(100, optimizer='two-stage')
model.fit(X_digits_cupy)
assert_array_equal(model.ranking[:3], digits_ranking[:3])
assert_array_almost_equal(model.gains[:3], digits_gains[:3], 4)
